ShiptoAverage

1
BMW Project

• Ship-to-Average
• by
• Matthias Pauli
• Thomas Drtil
• Claus Reeker
• Stefan Lier
• Christopher Vine
• Fernando Cruz

2
Plant Spartanburg

• 140,000 vehicles in 2004
• Over 6,000 part numbers for X5
• 70 option driven
• 40 of parts from Europe

3
Supply Chain
4
Challenges
5
Demand Variability
Standard Deviation 42/day
Mean Demand 78/day
) Data of engine 7781905-00, high runner
6
BMW policy Ship-to-forecast
Order Arrival
7
Inventory

• On-hand inventory with ship-to-forecast
• constant level?

) Data of engine 7781905-00, high runner
8
Forecast error

• Why try to chase the daily forecast?

9
Different forecasts
) Data of engine 7781905-00 , high runner
10
Approach Ship-to-average

• Dont ship to daily forecast
• Consider a longer forecast period instead
• Keep shipments constant, let the inventory swing

• Goals
• 1) Minimum impact on total avoidable costs
• 2) More stability for the supply chain

11
Basic Implementation

• Always ship average quantity!
• What happens to the inventory?

) Data of engine 7781905-00, high runner
12
How to control the inventory?
Deflate shipments Avg. forecast (x weeks)
deflation factor
Inflate shipments Avg. forecast (x weeks)
inflation factor
Max. Inventory Position
Inventory Position
(almost) constant shipment quantities !
Time
13
Which Part analyzed?

• Part
• Engine 7781905-00
• High runner
• Policy
• of weeks for average 3
• Max. Inventory Position 2509
• Inflation/deflation 1.8

14
Performance Overview

• How does ship-to-average perform for this engine

15
Shipment Comparison
ship-to-forecast
(shipment adjustment 66)
shipment quantity changes more than 10
compared to previous one
ship-to-average
(shipment adjustment 14)
Shipment adjustments happen in 14 of all
shipments
16
Whats next?

• Goals achieved! Optimized policy works.

• But how robust is the result?

• What are the trade-offs?

• How do the 3 parameter
• of weeks for average
• Max. inventory position
• Inflation/deflation factor
• influence the result?

17
Sensitivity Analysis

• of weeks for average

18
Sensitivity Analysis

• Max. Inventory Position

19
Sensitivity Analysis

• Inflation/deflation factor

20
Summary Table
21
Advantages

• Small cost reduction compared to current
  ship-to-forecast policy
• Less variation in order quantities
• Less bullwhip effect
• Easier operations for
• Spartanburg/ Wackersdorf/ upstream suppliers
• Facilitates negotiation with transportation
  partner

22
Limitations of the study

• Simulation vs. reality
• Restricted original data sets provided
• Small number of parts considered
• Constant shipment frequency assumed (once per
  week)

23
Recommendations

• Run pilot to check performance
• pick high runner with relatively stable demand
  over time
• Analyze larger set of parts
• Evaluate cost savings upstream
• Evaluate trade-off between higher savings and
  increasing expediting

24
QA